Hypoglycemia is a serious and common consequence of exogenous insulin therapy in type 1 diabetes mellitus (T1DM) and is a key barrier to the achievement of optimal glucose control. On average, hypoglycemia occurs 2 times weekly in T1DM and severe hypoglycemia, defined as an event where a patient requires the assistance of another person to administer carbohydrate or glucagon, has an incidence of 1.1 to 1.5 episodes per patient-year. A limitation of subcutaneous insulin therapy is that insulin is delivered into the peripheral circulation rather than physiologically into the portal circulation. likely consequence of this different route of delivery is increased exposure to hypoglycemia, as a portion of glycemic control is shifted away from the liver to skeletal muscle, where there is a larger mass of tissue to act as a glucose sink, where glucose is taken up from the circulation, even in hypoglycemic and euglycemic states, and where the response time to correct hypoglycemia is comparatively slower. The proposed research will test the hypothesis that delivery of insulin into the portal vein, as opposed to a peripheral vein, reduces hypoglycemic exposure in a canine model utilizing two approaches. In the first approach, we will apply a surgical model of selective hepatic insulin delivery (Aim 1). We will determine the extent to which equivalent overinsulinization brought about by insulin infusion into the portal rather than systemic circulation decreases resultant hypoglycemia. These experiments will help to determine how much restoring the balance of insulin distribution between the liver and other tissues in the body confers a protective effect against hypoglycemia. We will further examine how much the propensity for increased hypoglycemia associated with peripheral insulin delivery is exaggerated when the glucagon response to hypoglycemia is lost, as seen with T1DM (Aim 2). In the second approach, we will employ a novel hepatopreferential insulin analog to assess its ability to reproduce results seen with the surgical approach (Aim 3). These studies will set th stage for new approaches to insulin therapy that will permit more aggressive diabetes management while limiting hypoglycemia, a principal barrier to optimal control at present. If hypoglycemia is reduced by utilization of hepatopreferential approaches, a strong case could be made to either further develop insulin analogs that would take advantage of this approach or seek ways to deliver insulin intraportally. These studies will also provide a focus for the trainin of a new pediatric investigator in the physiology of glucose regulation using state of the art techniques in the laboratory of an established and highly successful mentor.